数据驱动下基于量子人工蜂群的 K 均值聚类算法优化

In this paper, the quantum artificial bee colony (QABC) algorithm is applied to the selection of category center points of K-means clustering,and the K-means clustering algorithm is optimized,which can effectively solve the problem of low clustering accuracy caused by randomly setting K-means center points. This paper sets the number of K-means clustering categories,and randomly sets several category centers. The artificial bee colony(ABC)algorithm is used to optimize the category center. The bee colony is constructed and clustered according to the sample points,and the individual positions of the bee colony are represented by quantum bits. This paper takse the reciprocal of the distance between the sample point and the center point as the fitness of ABC algorithm,and defines the individuals with higher fitness value as honey sources. By leading the bees through coarse-grained traversal in the range of motion and following the bees through fine-grained exploration,individuals with high fitness are searched continuously,and honey sources are updated until the ABC algorithm is stable,and the optimal honey source position is determined as the cluster center. K-means clustering is performed by using the clustering center optimized by ABC. The experimental results show that by reasonably setting the ABC search boundary and introducing the quantum representation of the position of the bee colony,the accuracy of ABC's search for the cluster center can be effectively enhanced. Compared with the common clustering algorithms,the clustering performance of QABC+K-means algorithm is better.